Convergence of Nekrasov instanton sum with adjoint… — Plain-Language Explanation

✨

This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine you are trying to bake the perfect cake, but the recipe is written in a language you don't quite understand. The recipe calls for an infinite number of ingredients, added one by one. The question isn't just "does the cake taste good?" but "does the cake even exist?" If you keep adding ingredients forever, does the mixture eventually settle into a stable, delicious cake, or does it explode into a chaotic, inedible mess?

This is the story of Bruno Le Floch's paper, which tackles a very similar problem in the world of theoretical physics.

The Setting: The Infinite Cake Recipe

In the universe of quantum physics, specifically in a theory called $N=2^*$ gauge theory, physicists use a mathematical tool called the Nekrasov Instanton Partition Function.

Think of this function as a giant, infinite recipe for calculating the energy of a quantum system.

The Ingredients: The "ingredients" are called instantons. These are tiny, fleeting quantum fluctuations.
The Counting Parameter ( $q$ ): This is like the "oven temperature" or the "amount of flour." It's a number that tells you how many instantons you are considering.
The Recipe Steps: The recipe says, "Take 1 instanton, then 2, then 3, then 4... all the way to infinity." Each step adds a new term to the sum.

The big question is: If we add up all these infinite terms, do we get a finite, sensible number, or does the sum blow up?

The Problem: When Does the Cake Explode?

Usually, in physics, these infinite sums are just "asymptotic." That means they work for a few terms, but if you keep going, they diverge (explode). However, physicists suspect that for this specific type of theory, the sum should converge (settle down) as long as the "oven temperature" ( $q$ ) isn't too high. Specifically, they expect it to work if $|q| < 1$ .

But there's a catch. The behavior of the sum depends heavily on a hidden parameter called $b^2$ (which is a ratio of two other numbers, $\epsilon_1$ and $\epsilon_2$ ). Think of $b^2$ as the texture of the flour.

If the flour is "smooth" (generic), the cake bakes perfectly.
If the flour is "weird" (specific irrational numbers), the cake might be unstable.
If the flour is "broken" (rational numbers), the recipe might be impossible to follow because you hit a division by zero.

The Discovery: The "Flour Texture" Matters

Bruno Le Floch's paper is a rigorous proof that answers: "Exactly when does this infinite recipe work?"

He breaks the problem down into three scenarios based on the "texture" of $b^2$ :

1. The Smooth Case ( $b^2$ is not a real number)

If $b^2$ is a complex number (imagine a flavor that doesn't exist in our 2D world of real numbers), the recipe is perfectly safe.

The Result: The sum converges for any $|q| < 1$ .
The Analogy: It's like baking with a standard, high-quality flour. No matter how much you mix, the batter stays stable up to the limit.

2. The "Almost Rational" Case ( $b^2$ is a positive irrational number)

This is the tricky part. Some irrational numbers are "close" to being fractions (like 3.14159... is close to 22/7).

The "Bad" Flour: If $b^2$ is an irrational number that is extremely well-approximated by fractions (mathematicians call these Liouville numbers), the recipe fails. The sum explodes for any amount of flour ( $q \neq 0$ ).
The "Okay" Flour: If $b^2$ is irrational but not too close to fractions (a property related to Brjuno numbers), the recipe works, but the "safe zone" for $q$ shrinks. The closer $b^2$ is to a fraction, the smaller the safe oven temperature becomes.
The Analogy: Imagine the flour is slightly damp. If it's too damp (super-approximable), the cake collapses immediately. If it's just a little damp, you can still bake, but you have to be very careful with the temperature.

3. The Broken Case ( $b^2$ is a rational number)

If $b^2$ is a simple fraction (like 1/2 or 3/4), the recipe contains division by zero.

The Result: The sum is "ill-defined." Some terms are infinite.
The Analogy: The recipe asks you to divide a cake into zero pieces. It's a logical impossibility.
The Twist: The paper notes that while individual terms blow up, the total sum might still make sense if you group the terms differently (like canceling out the infinite parts). But as written, the raw recipe is broken.

Why Should We Care? (The AGT Connection)

You might ask, "Who cares about an infinite cake recipe?"

The paper connects this to the AGT Correspondence, a famous "Rosetta Stone" in physics that translates between:

4D Quantum Physics (The cake recipe).
2D Conformal Field Theory (A different kind of math describing shapes and vibrations).

By proving when the cake recipe converges, Le Floch also proves when the Conformal Blocks (the 2D translation) converge.

The Big Win: He proves that for a huge range of parameters (specifically when the "central charge" of the theory is not a specific large number), these 2D mathematical objects are well-behaved and converge inside the unit disk ( $|q| < 1$ ).

The "Secret Sauce" of the Proof

How did he prove this?
He didn't just guess. He looked at the denominators of the fractions in the recipe.

In the "smooth" cases, the denominators stay large enough to keep the fractions small.
In the "weird" irrational cases, he had to use advanced number theory (like Continued Fractions) to measure exactly how close the denominators get to zero.
He showed that if the denominators get too close to zero too fast (the "super-approximable" case), the terms grow so fast that the sum explodes.

Summary in a Nutshell

Bruno Le Floch has mapped out the safety zone for a complex quantum physics formula.

If the parameters are generic: The formula works perfectly up to a limit of 1.
If the parameters are "weirdly" irrational: The safety zone shrinks depending on how "fraction-like" the number is.
If the parameters are rational: The formula breaks down (unless you rearrange the terms).

This is a foundational result. It tells physicists exactly when they can trust their calculations and when they need to be careful, bridging the gap between abstract number theory and the fundamental laws of the universe.

1. Problem Statement

The paper addresses the mathematical convergence properties of the Nekrasov instanton partition function for the 4d $\mathcal{N}=2^*$ $U(N)$ gauge theory. This theory is a mass deformation of 4d $\mathcal{N}=4$ super-Yang-Mills theory, consisting of a $U(N)$ vector multiplet coupled to an adjoint hypermultiplet of mass $m$ .

The partition function is expressed as a generating series over colored partitions (tuples of Young diagrams $\vec{Y}$ ):
$Z_{\text{inst}}(m, a; q) = \sum_{\vec{Y}} q^{|\vec{Y}|} Z_{\vec{Y}}(m, a)$
where $q$ is the instanton counting parameter, $a$ represents Coulomb branch parameters, and $\epsilon_1, \epsilon_2$ are equivariant parameters of the Omega background.

The Core Question: While the series is known to be related to 2D Conformal Field Theory (CFT) conformal blocks via the AGT correspondence, it is not immediately obvious whether this series converges. In many quantum field theories, such expansions are merely asymptotic. The author seeks to determine the absolute convergence radius ( $R_{\text{abs}}$ ) of this series, specifically investigating how $R_{\text{abs}}$ depends on the ratio $b^2 = \epsilon_1/\epsilon_2$ and the genericity of the parameters $m$ and $a$ .

2. Methodology

The author employs a rigorous analytic approach based on bounding the magnitude of the coefficients $Z_{\vec{Y}}$ for large partition sizes $|\vec{Y}| = k$ .

Explicit Formulas: The coefficients $Z_{\vec{Y}}$ are products of rational functions involving the box contents of Young diagrams. The analysis focuses on the denominators of these rational functions, which take the form of linear combinations of $\epsilon_1, \epsilon_2$ and the parameters $a, m$ .
Lattice Analysis: The convergence depends critically on how close the values $- \epsilon_1(\mu'_j - i) + \epsilon_2(\lambda_i - j)$ come to zero. This relates to the distribution of the lattice $\epsilon_1 \mathbb{Z} + \epsilon_2 \mathbb{Z}$ relative to the parameters.
Case Distinction: The analysis is split based on the nature of the ratio $b^2 = \epsilon_1/\epsilon_2$ $b^{2} = ϵ_{1} / ϵ_{2}$ :
1. Non-real $b^2$ : The lattice is discrete.
2. Negative real $b^2$ : The lattice is discrete but lies on a line.
3. Positive irrational $b^2$ : The lattice is dense in a line; convergence depends on Diophantine approximation properties (how well $b^2$ is approximated by rationals).
4. Positive rational $b^2$ : The lattice contains zero, leading to potential singularities.
Combinatorial Bounds: The author derives bounds on the number of boxes in a Young diagram that share specific "content" values. A key insight is that the number of boxes with a given content value grows sub-linearly (specifically as $O(\sqrt{|\vec{Y}|})$ ), which prevents the product of terms from diverging too rapidly.
Diophantine Tools: For $b^2 > 0$ , the paper utilizes concepts from number theory, specifically continued fractions and a variant of Brjuno numbers (specifically the "exponential type" $B_{\text{sup}}(b^2)$ ), to quantify how well $b^2$ can be approximated by rationals.

3. Key Contributions and Results

The main result is Theorem 1.1, which determines the absolute convergence radius $R_{\text{abs}}$ under generic assumptions (parameters avoiding the closure of the lattice $\epsilon_1 \mathbb{Z} + \epsilon_2 \mathbb{Z}$ ).

A. Non-Generic Cases (Rational $b^2 > 0$ )

Result: If $b^2$ is a positive rational number, the sum is ill-defined (meaningless) because specific terms in the sum become singular (poles).
Nuance: While the individual terms diverge, the author notes that in the full partition function (summed over all tuples), these singularities might be removable. However, the absolute sum (sum of absolute values) diverges.

B. Generic Cases: $b^2 \in \mathbb{C} \setminus [0, +\infty)$

Result: For non-real $b^2$ or negative real $b^2$ , the absolute convergence radius is optimal:
$R_{\text{abs}} = 1$
Significance: This confirms the physical expectation that the series converges within the unit disk for these regimes, independent of the specific values of $m$ and $a$ (provided they are generic).

C. Positive Irrational $b^2$

This is the most complex regime, where the lattice is dense. The convergence depends on the exponential type $B_{\text{sup}}(b^2)$ , defined as:
$B_{\text{sup}}(b^2) = \limsup_{n \to \infty} \frac{\log q_{n+1}}{q_n}$
where $p_n/q_n$ are the convergents of the continued fraction of $b^2$ .

Finite Exponential Type: If $b^2$ is "not too well" approximated by rationals (finite $B_{\text{sup}}$ ), the radius is strictly positive but less than 1:
$A_1 e^{-8 B_{\text{sup}}(b^2)/\max(1, b^2)} \leq R_{\text{abs}} \leq \min(1, A_2 e^{-B_{\text{sup}}(b^2)/(1+b^2)})$
Here, $A_1, A_2$ are continuous functions of the physical parameters.
Infinite Exponential Type: If $b^2$ is "super-exponentially well approximable" (e.g., Liouville numbers, where $B_{\text{sup}} = \infty$ ), then:
$R_{\text{abs}} = 0$
The series diverges for any $q \neq 0$ .

4. Implications for Conformal Field Theory (AGT Correspondence)

The paper translates these gauge theory results to the convergence of torus one-point conformal blocks for the Virasoro and $\mathcal{W}_N$ algebras via the AGT correspondence.

Virasoro Algebra ( $N=2$ ): The result establishes convergence for the conformal blocks within the unit disk $|q| < 1$ for central charges $c_{\text{Vir}} \in \mathbb{C} \setminus [25, +\infty)$ .
General $\mathcal{W}_N$ : Convergence is established for generic parameters when the central charge lies in a specific complex domain excluding a real interval.
Distinction: The paper clarifies that for $b^2 > 0$ (corresponding to $c > 25$ ), convergence is not guaranteed for all parameters; it fails for Liouville-type parameters (infinite exponential type) and is ill-defined for rational parameters without careful regularization.

5. Significance

Mathematical Rigor: This work provides the first rigorous proof of the convergence radius for the Nekrasov partition function in the presence of adjoint matter, moving beyond the assumption that these series are merely asymptotic.
Arithmetic Dependence: It reveals a deep connection between the convergence of physical observables in QFT and the arithmetic properties of the coupling constants (specifically the Diophantine properties of $b^2$ ). This mirrors phenomena seen in KAM theory and small divisor problems in dynamical systems.
Optimality: The proof establishes that $R_{\text{abs}}=1$ is the optimal radius for generic non-positive $b^2$ , confirming long-standing expectations in the physics community.
Technique: The method of bounding the "box content" distribution and relating it to the density of the lattice offers a new toolkit for analyzing similar partition functions in 5d theories or other supersymmetric gauge theories.

In summary, the paper resolves the convergence question for the 4d $\mathcal{N}=2^*$ theory, showing that while the series generally converges in the unit disk, the absolute convergence radius can shrink to zero or become undefined depending on the arithmetic nature of the equivariant parameter ratio $b^2$ .

Convergence of Nekrasov instanton sum with adjoint matter